際際滷

際際滷Share a Scribd company logo
Are Twitter researchers going to
hell?
29-05-2015 1
Will I go to hell ?
29-05-2015 揃 2
29-05-2015 揃 3
29-05-2015 揃 4
}Lust: In the second circle
of Hell are those overcome
by lust. Dante condemns
these "carnal malefactors"
for letting their appetites
sway their reason.
29-05-2015 揃 5
- Social media users are not a representative
sample
- Facebook: 53% (of US population)
- LinkedIn: 19%
- Twitter: 13%
(PEW data 2013  Hargittai 2015)
29-05-2015 揃 6
}Is Twitter a good proxy
of social life?
U
S
T
w
i
t
t
e
r
This is where it goes
29-05-2015 揃 7
Temporal Patterns of Happiness and
Information in a Global Social Network:
Hedonometrics and Twitter
http://dx.plos.org/10.1371/journal.pone.0026752
- Twitter users are largely
unknown
- There are no reasons to
expect any trickle down
effect from Twitter to
society
29-05-2015 揃 8
}Is Twitter a good proxy
of social life?
29-05-2015 揃 9
}Heresy: In the sixth circle,
Heretics, such as Epicureans
("the soul dies with the
body") are trapped in flaming
tombs.
- Research questions are
defined by the possibilities
offered through the API
- Data monetization has been
a clear strategy behind the
evolution of Twitter API
29-05-2015 揃 10
}Dealing with
privately-own data
1)
- API (stream/rest) based acquisition do not provide
representative samples for big events (Morstatter et al. 2013)
- Using Twitter data providers can be extremely expensive
(for average research budget in social sciences /
humanities)
[e.g. Rule Text: eurovision
Start Date: 05/18/2015
End Date: 05/25/2015
Estimated Activities: 2,042,000
The total cost of accepting this job is $580.00. ]
29-05-2015 揃 11
}Dealing with
privately-own data
2)
29-05-2015 揃 12
}
Fraud: Sorcerers,
astrologers, and false
prophets here have their
heads twisted around on
their bodies backward, so
that they "found it necessary
to walk backward, / because
they could not see ahead of
them..
Hard methods for soft data
29-05-2015 揃 13
Quantitative methods often
assume the stability of the
research field.
Can we still cite/use Twitter
data published in 2007 or 2008
?
29-05-2015 揃 14
Connect with your friends  and
other fascinating people. Get in-the-
moment updates on the things that
interest you. And watch events
unfold, in real time, from every
angle.
Practices > structure
29-05-2015 揃 15
Whats happening?What are you doing?
29-05-2015 揃 16
Are we doing the research we want to do
or the research we can do ? And what
are the  systemic  forces affectingt this
process?
One more thing
29-05-2015 揃 17
29-05-2015 揃 18
Thank you

More Related Content

Going to hell_slideshare

  • 1. Are Twitter researchers going to hell? 29-05-2015 1
  • 2. Will I go to hell ? 29-05-2015 揃 2
  • 4. 29-05-2015 揃 4 }Lust: In the second circle of Hell are those overcome by lust. Dante condemns these "carnal malefactors" for letting their appetites sway their reason.
  • 6. - Social media users are not a representative sample - Facebook: 53% (of US population) - LinkedIn: 19% - Twitter: 13% (PEW data 2013 Hargittai 2015) 29-05-2015 揃 6 }Is Twitter a good proxy of social life? U S T w i t t e r
  • 7. This is where it goes 29-05-2015 揃 7 Temporal Patterns of Happiness and Information in a Global Social Network: Hedonometrics and Twitter http://dx.plos.org/10.1371/journal.pone.0026752
  • 8. - Twitter users are largely unknown - There are no reasons to expect any trickle down effect from Twitter to society 29-05-2015 揃 8 }Is Twitter a good proxy of social life?
  • 9. 29-05-2015 揃 9 }Heresy: In the sixth circle, Heretics, such as Epicureans ("the soul dies with the body") are trapped in flaming tombs.
  • 10. - Research questions are defined by the possibilities offered through the API - Data monetization has been a clear strategy behind the evolution of Twitter API 29-05-2015 揃 10 }Dealing with privately-own data 1)
  • 11. - API (stream/rest) based acquisition do not provide representative samples for big events (Morstatter et al. 2013) - Using Twitter data providers can be extremely expensive (for average research budget in social sciences / humanities) [e.g. Rule Text: eurovision Start Date: 05/18/2015 End Date: 05/25/2015 Estimated Activities: 2,042,000 The total cost of accepting this job is $580.00. ] 29-05-2015 揃 11 }Dealing with privately-own data 2)
  • 12. 29-05-2015 揃 12 } Fraud: Sorcerers, astrologers, and false prophets here have their heads twisted around on their bodies backward, so that they "found it necessary to walk backward, / because they could not see ahead of them..
  • 13. Hard methods for soft data 29-05-2015 揃 13
  • 14. Quantitative methods often assume the stability of the research field. Can we still cite/use Twitter data published in 2007 or 2008 ? 29-05-2015 揃 14
  • 15. Connect with your friends and other fascinating people. Get in-the- moment updates on the things that interest you. And watch events unfold, in real time, from every angle. Practices > structure 29-05-2015 揃 15 Whats happening?What are you doing?
  • 17. Are we doing the research we want to do or the research we can do ? And what are the systemic forces affectingt this process? One more thing 29-05-2015 揃 17